Pullback Trading Strategy: Buying the Dip in an Uptrend, Backtested Live

"Buy the dip" is half a strategy. Everyone knows the first half, the stock you like goes down a few percent and you're tempted. The half that separates a pullback system from wishful averaging-down is the boring part: which dips qualify, where exactly you get out if you're wrong, and how much you risk finding out.

This is the complete version of the trade, with every rule mechanical enough to backtest, and backtested, live, further down the page.

The idea

Strong stocks don't go up in a line. They advance, exhale back to a rising moving average, then advance again. A pullback trader ignores the advances entirely and only shops during the exhales: you're buying a stock that has already proven it's in demand, at a discount to where the buyers last were, close to a level where they've shown up before.

That framing does two useful things. It forces you onto strong stocks (there's no pullback without an uptrend first), and it gives you a natural line in the sand, if the "dip" keeps going and takes out the level, the premise is gone and you leave cheap.

The rules

The version our backtester runs, known to swing traders as the EMA bounce:

1. Trend filter. Price must be above its 200-day average. No uptrend, no trade, a pullback in a downtrend is just a downtrend.
2. The pullback. Price dips to the 50-day EMA, the classic "institutional shopping level" for a trending large cap.
3. The entry. You don't buy the touch, falling knives touch averages too. You buy when price reclaims the average, closing back above it. The bounce is the confirmation.
4. The stop. An ATR-based stop below the entry, 1.5× the Average True Range, so the exit scales with how much the stock normally wiggles instead of using a fixed percent that's too tight for a volatile name and too loose for a calm one. This distance is your 1R.
5. The target and the free roll. Take profit at 2R, twice the risk. And once the trade is up 1R, move the stop to breakeven: from there the worst case is a scratch.
6. The size. Risk 1% of the account per trade. Position size falls out of the stop distance; you never size by how much cash you feel like spending. (Why this matters more than the setup itself: see our swing-trading playbook.)

The live backtest

Here is that exact rule set on a liquid large cap: 200-day trend filter, 50-EMA bounce entry, 1.5 ATR stop, 2R target, breakeven at 1R, 1% risk, fees and slippage on. Run live by our engine. The card recomputes as new prices arrive; clicking it opens the same run in the tool:

Trend pullback on AAPLloading live result…Past result of the mechanical rule, fees on, not a prediction.Open this exact setup in the backtester

The number to look at first is expectancy, the average result per trade in units of risk. A 2R-target system loses often by design (the stop is closer than the target, and breakeven exits turn near-misses into scratches), so the win rate will look unimpressive. That's fine. +0.2R or better per trade across a decent sample is a real edge; what kills pullback systems is a slightly negative expectancy that a trader never measures because the wins feel frequent enough.

Also worth checking:

• Time in market. A filtered pullback system sits in cash a lot. The benchmark buy-and-hold line is compounding the whole time. That's the honest comparison, and on a strong stock it's hard to beat.
• The same rules on a different ticker. This is the most instructive experiment in the tool: the identical system is often good on one trending name and a steady bleed on a choppier one. The edge lives in what you trade it on as much as in the rules.

The same rules across 22 stocks

The card above is one ticker, and the most common objection to any single-name backtest is that it is one ticker. So we ran it across 22 US large caps and sector ETFs over the same window, changing one thing at a time.

The exits matter more than the sizing, and it is not close. Raising risk from 1% to full size moves the median from +2.6% to +6.3%. Removing the ATR stop and the 2R target moves it from +6.3% to +27.7%. The rule that decides when you get out is worth several times the rule that decides how much you buy, which is the reverse of where a beginner's attention usually goes and the reverse of what the first version of this section claimed.

That is not an argument for trading without a stop. The stop is what makes the drawdown column read -2.6% instead of -21.1%, and a 2R target is a deliberate decision to cap the winners that pay for a low win rate. It is an argument that the target is the expensive half of the bargain: a 2R cap in a market that ran a long way is what turned +27.7% into +6.3%.

Two honest cautions about reading this table, because both numbers in column one look better and worse than they are:

• +2.6% is not a poor return on the money at work. Risking 1% per trade leaves the account roughly 92% in cash, so this is a small return on a mostly idle account, not a small return on the trades. Comparing it to a fully invested 57.3% compares two different things, and so does "0 of 22" — at 7.7% exposure that outcome is close to arithmetically forced and settles nothing.
• The -2.6% drawdown is mostly the cash, not the stop. 7.7% exposure against a -33.8% benchmark drawdown is about -2.6% on its own. The stop is doing real work, but this column is not the evidence for it; the full-size column, at -8.2% against the same -33.8%, is.

Variations

• Faster pullbacks (20-EMA), slower trends (100/150-day): shorter exhales, more trades, noisier. All one field each in the trend pullback backtester.
• Ride the trend instead of taking 2R: drop the target, keep a trailing stop, and hold as long as the trend holds. Fewer, bigger winners; more round trips that give everything back. We covered this variant, the trend retracement, in the playbook.
• Confirmation at the level: many traders demand a reversal candle at the average, a bullish pin bar, an engulfing bar, before the entry. The backtester trades the mechanical reclaim; the candle read is judgement layered on top.

The honest verdict

Pullback trading is probably the most defensible setup a beginner can learn: it forces you into strong stocks, it has a built-in "I'm wrong" point, and risk-based sizing caps the damage of any one mistake. It is also relentlessly humbling in practice, the filter keeps you out for months, the breakeven rule scratches trades that then run without you, and a strong stock's buy-and-hold line is a hard benchmark to beat.

Which of those dominates for the ticker you'd trade is exactly one click away. The run above is live: swap the ticker, loosen the stop, drop the filter, and watch what each change does to the expectancy before you ever risk a euro on it.

Open the pullback setup in the backtester →